Multiobjective optimization of a journal bearing using the Pareto optimality concept

Abstract

This paper deals with multiobjective optimization (minimization of temperature 
rise, minimization of oil feed flow, minimization of power loss) and discrete design 
variables (lubricant viscosity, radial clearance, length-diameter ratio) related to 
the design of a journal bearing. The aim of this paper is to show the sensitivity of 
radial clearance, oil viscosity and length-diameter ratio on bearing power loss and 
oil flow. The paper attempts to remove the difficulty of selecting the weighting factor 
in multiobjective journal bearing optimization problems.

The optimization model is based on a hybrid solution scheme for journal bearings. 
The objective of bearing optimization is to minimize a linear combination (weighted sum) 
of power loss and oil flow. Constraints are imposed on the maximum fluid film 
pressure, minimum film thickness, maximum temperature rise and critical speed. A generic 
algorithm is used as the optimization tool with a penalty to those solutions that violate 
constraints. The three-level Taguchi method is used to measure the effect of design 
variables on the optimum solution. The Pareto-optimal set approach is adapted to segregate 
the superior design vectors.

The results of power loss, oil flow, temperature rise, maximum pressure and minimum film 
thickness are presented in tabular and graphical form for comparative study. Three journal 
bearing design examples are included, demonstrating the effectiveness of the proposed method.